Downloads · 30 days
0
Qwen3.8-3.6-27B-Splash
audiohacking/Qwen3.8-3.6-27B-Splash
Qwen3.8-3.6-27B-Splash is a image-text-to-text model from audiohacking. Use it for the image-text-to-text task on the model card, and read the license before you ship it in a product. It is set up for mlx. The card lists the license as apache-2.0.
Splash 1.0 schema-3 package of the JetBrains Qwen3.8 × Qwen3.6 27B blend.
Access
Public
Updated Sep 20, 2026
Repo size
28.5 GB
Likes
0
Public
Repo makeup
Click a slice to open those files.
.bin17.4 GB · 100%
At a glance
- Task
- Image-Text-to-Text
- Library
- mlx
- License
- apache-2.0
- Access
- Public
- Created
- Sep 20, 2026
- Updated
- Sep 20, 2026
- SHA
- b80ce931
Try a prompt
Base models
Model card
From the Hugging Face model README
Qwen3.8-3.6-27B-Splash
Splash 1.0 schema-3 package of the JetBrains Qwen3.8 × Qwen3.6 27B blend.
splash serve --model audiohacking/Qwen3.8-3.6-27B-Splash
What is packed
| Section | Source |
|---|---|
target/ | Packed from JetBrains/Qwen3.8-3.6-27B-blend-MLX-4bit (affine 4-bit, group 64) |
vision/ | Packed from the same blend vision tower |
tokenizer/ | Blend tokenizer + Splash chat-template tweak (later system messages rendered in place) |
draft/ | Packed DFlash2 copied from incoai/Qwen3.8-27B-Splash. Splash requires this layout; drafted tokens are verified by the blend target |
This is a Hugging Face model repo. Splash will not load the BF16 bucket audiohacking/Qwen3.8-3.6-27B-blend.
JetBrains Qwen3.8-3.6-27B-blend-MTP-MLX-4bit is an mlx-vlm drafter, not a Splash draft/ replacement.
- Task
- Image-Text-to-Text
- Library
- mlx
- License
- apache-2.0
- Created
- Sep 20, 2026
- Updated
- Sep 20, 2026